CN1685750A - A method and unit for eliminating multi-user interference - Google Patents

Abstract

The invention discloses a method and a unit for eliminating multi-user interference, wherein the method mainly comprises the following steps: performing a de-multiplexing spread spectrum operation on the input baseband signal by using a de-spreading unit; then, the de-spread signal is continuously subjected to Walsh de-spread of the sub-channel, and the signals of each channel of the user are separated; multiplying the separated bit streams of each channel with a conjugate signal of an output estimation value of a pilot frequency estimator A to counteract the influence of multipath fading; the multiplied bit streams are sent to a decision device for decision; the invention uses another two pilot estimators B, C and the number of users, channel type, service type provided by the system to construct the threshold needed by the decision, the invention can control and process the signal decision and signal recovery more accurately, and increase the accuracy of interference elimination, thus improving the system performance.

Description

A multi-user interference cancellation method and unit Technical field

The present invention relates to multi-user detection in a code division multiple access communication system, and more specifically refers to a multi-user interference elimination method and unit. Background technique

Code Division Multiple Access (CDMA) is one of the multiple access modulation techniques commonly used in mobile communications. Other multiple access modulation technologies include time division multiple access (TDMA) and frequency division multiple access (FDMA). Compared with other multiple access modulation technologies, code division multiple access technology has the advantages of high frequency band utilization and large system capacity.

In a CDMA communication system, user signals overlap in time domain and frequency domain, and are distinguished from each other only by spreading codes. If the spreading codes are completely orthogonal, the signals of each user can be completely recovered by using a correlator or a matched filter. However, in an actual system, users are not synchronized and signals of different users arrive at the receiver with different delays, so it is difficult to find a certain spreading code sequence to make the signals of all users in all possible relative delay ranges Inner Orthogonal. Since the spreading code set is not completely orthogonal, there is interference between different users in the channel and different multipaths of the same user, that is, multiple access interference (or multi-user interference). Traditional receivers use correlators (or matched filters) for demodulation, and the interference between users cannot be removed during the correlation process. For a certain user, all signals of other users are regarded as noise. Therefore, as the number of users in the system gradually increases, the interference between users also increases. When these interferences accumulate to a certain extent and exceed the minimum signal-to-noise ratio required for system demodulation, the system cannot access more User. Therefore, the CDMA system is an interference-limited system. To solve the limited interference problem of the CDMA system to some extent, it is necessary to reduce the impact of multi-user interference. In fact, different from system thermal noise, multiple access interference (MAI: Multiple Access Interference) is theoretically estimable and reproducible. Therefore, it is possible to synthesize the useful information of each user and use a certain signal processing method to reduce the multiple access interference in the received signal, which is the goal to be achieved by the multi-user detection technology. Effective multi-user detection technology can improve system capacity, increase system coverage radius, and alleviate the interference limitation problem of CDMA.

At the same time, interference cancellation (multi-user detection) technology can also effectively alleviate the "near-far effect" on impact on system performance. Due to the different distances between the mobile station and the base station and the influence of wireless channel fading, the signal power received by the base station from each mobile station will be different, and users with strong-power signals will cause great interference to users with weak-power signals. Users of weak power signals experience degraded or even non-functional functionality. Using power control can make the power of all mobile stations received by the base station approximately equal, and alleviate the "near-far effect" to a certain extent. However, power control also has some disadvantages, such as the need to occupy channels to transmit power control information, control lag, performance and mobile user rate correlation, etc., and power control cannot solve the problem of multiple access interference limiting system capacity. The starting point of power control is to control the interference at an acceptable level. Unlike this, the use of interference cancellation technology is to remove the interference between users to the greatest extent, which fundamentally eliminates the cause of the "near-far effect". , so it can effectively alleviate the influence of the "near-far effect" and reduce the performance requirements of the system on power control.

Usually mobile communication is faced with a time-varying multipath fading environment. In such an environment, the transmitted signal will arrive at the base station with different delays after propagating through different paths, and since the spreading codes are not completely orthogonal, there is also mutual interference similar to MAI between different multipath signals of the same user . In the traditional CDMA receiver design, Rake branches are used to demodulate the strongest several-path signals of each user separately, and then combined at the maximum ratio, and diversity is used to overcome the influence of multi-path fading. However, in the process of demodulating each path, other multipaths are still regarded as noise, and the information contained therein is not fully utilized. Therefore, in a multipath propagation environment, only simple diversity combining reception cannot provide the effect of overcoming multiple access interference and multipath influence at the same time. If the influence of multipath is considered in the multi-user detection algorithm, it is obvious that effective information can be fully utilized, thereby further improving system performance.

Multi-user detection receivers can be divided into linear multi-user receivers and nonlinear multi-user receivers according to whether their structure has feedback.

The linear multiuser detector equates the multiple access interference in a multiuser communication environment to a channel transmission response matrix, and the transmission matrix is related to each user's spreading sequence and the relative delay between the sequences. If we can get the inverse matrix of the channel transmission matrix, we can pass the multi-user signal through the output of k (k is the number of users) matched filters, and then perform an inverse operation through this inverse matrix, so as to eliminate the difference between each user equivalently. Correlation between, so as to achieve the purpose of eliminating multiple access interference. However, since this method needs to know the precise phase information between the spreading codes, and also needs to calculate the inverse matrix of the correlation matrix at any time according to the change of the actual user or the change of the environment, the algorithm is complicated, the calculation amount is large, and it is not easy to realize in real time. Another important class of multiuser detectors is called nonlinear multiuser detectors (also called subtractive interference cancellation detectors). Its basic principle is to independently estimate the MAI information from each user at the receiving end, and then subtract part or all of the MAI of the corresponding user from the total received signal to obtain the interference-free signal of each user, and then use demodulation by conventional receivers. This kind of feedback-based detector is usually implemented in a multi-stage manner, and it is expected to maximize the interference cancellation through multi-stage feedback, so as to obtain better demodulation performance. It can be seen from the analysis of the linear multiuser detector that the matrix operation is mainly used in the linear multiuser detector, and the calculation is relatively complicated, which is not conducive to the realization of hardware. From an implementation point of view, nonlinear multiuser detectors are more efficient.

Nonlinear multiuser detectors can be implemented in serial or parallel architectures. A detector with a serial structure generally needs to perform power sorting on the input signals, first perform interference cancellation on users with stronger power in order, and then use the signal after interference cancellation as input, and do the same processing on signals with weaker power. When the effect of power control is not obvious or there is lag, the performance of serial processing is better than that of parallel processing, but it will introduce a processing delay proportional to the number of users; the power between users is relatively balanced or the system requires processing delay When it is higher, a parallel structure is usually used. No matter which structure is adopted, multi-stage processing is generally performed to obtain a better interference elimination effect.

The performance of multiuser detectors is determined by the ICU (Interference Cancellation Unit) regardless of the serial or parallel structure. ICU usually includes two parts: signal judgment and signal recovery. The goal of signal judgment and signal restoration is to accurately reconstruct the data of each user at the receiving end, so that the interference of other users can be removed during interference elimination, and at the same time, it is also necessary to ensure that no additional errors are introduced due to signal reconstruction errors during interference elimination. interference. The accuracy of signal judgment and signal recovery will directly affect the signal reconstruction capability of the ICU, thereby determining the performance of the entire detector. Therefore, how to accurately perform signal judgment and signal recovery has become a key factor to improve the performance of the detector.

Current research on multi-stage interference cancellation can be roughly divided into two categories: non-decision interference cancellation and decision interference cancellation (also called soft-decision interference cancellation and hard-decision interference cancellation). Non-decision interference cancellation directly uses the output of the correlation receiver to generate the interference recovery signal, and does not need to estimate the channel parameters during the processing. Its algorithm is relatively simple, and the gain of interference cancellation can be obtained in the white noise channel; but in the Rayleigh channel, if the influence of multipath fading is not considered, the signal obtained without RAKE processing cannot overcome the influence of fading, and the interference recovery If there is no reconstruction of the multi-channel If there is no path signal, then the restored signal obtained from it will be quite different from the actual received signal. After the interference is eliminated, interference may be introduced, which will directly lead to the performance degradation of the system in the Rayleigh environment. Judgment interference elimination uses the RAKE combined signal for judgment, and restores the multipath signal during the reconstruction process, which can effectively remove multi-user interference and multipath interference. NEC's patent 6081516, "Multiuser Receiving Device For Use In A CDMA System" (; "Multiuser Receiver For Use In A CDMA System"), first makes a hard decision on the data combined by RAKE, and then uses path estimation to restore multiple path signal, and then perform interference cancellation, which is suitable for Rayleigh fading environment. However, due to the lack of corresponding processing according to the characteristics of the received signal in the judgment, when the amplitude of the received signal of a certain user or a certain path is small, the reliability of using this signal for judgment and recovery is relatively low. In this case, due to The interference recovery signal is inaccurate and will instead introduce interference during the interference elimination process.

Paper "Third Generation Mobile Radio Systems Using Wideband CDMA Technology and Interference Canceller for Its Base Station" (JE FUJITSU Sci. Tech.J.,34,l,pp.50-57(September 1998)), namely "Using Wideband CDMA Technology Some techniques are adopted in the third generation wireless mobile system and interference cancellation at the base station to improve the accuracy of signal judgment and signal recovery. Firstly, mixed decision is used in signal decision, specifically, a threshold is set according to the energy of path estimation, if the energy of RAKE-combined signal is greater than the threshold, the decision is +1 or -1; if it is less than the threshold, the decision is a Value for threshold normalization (less than 1 ). It can be understood that when it is greater than the threshold, the reliability of the received signal is high, and it can be considered that the judgment is accurate; but when it is less than the threshold, the reliability of the received signal is poor, and it is very likely to misjudge, even if the judgment is a relative amplitude If the value is small, if the judgment is wrong, the interference will still be enhanced when the interference is eliminated. In this way, in multi-stage processing, the accumulation of errors will cause a rapid decline in performance.

Paper "Successive Interference Cancellation for Multiuser Asynchronous DS/CDMA Detectors in Multipath Fading Links" (see IEEE TRANSACTIONS ON COMMUNICATIONS, VOL.46, NO.3, MARCH 1998), § "Used in Multipath Fading Environment for Asynchronous DS/CDMA The algorithm of threshold judgment is also adopted in "Serial Interference Elimination Technology of Receiver". Different from the previous paper, no decision is made when the combined signal energy is less than the threshold. In addition, the calculation of the threshold in this paper is calculated according to the received signal energy variance of the user to be demodulated, the energy variance of the interfering user, and the noise energy variance. But in the actual system, to get the quasi-variance of these three kinds of energy It is not a simple process to determine the value, and if the threshold value involving the three parameters is to be calculated according to the change of the environment, more complicated calculation is required. Contents of the invention

The purpose of the present invention is to propose a multi-user interference cancellation method and unit for the defects of inaccurate signal judgment and signal recovery in the existing interference cancellation unit.

In order to achieve the above object, the present invention adopts the following technical solutions,

The method for multi-user interference cancellation of the present invention includes the following steps-a, using a despreading unit to perform decomplex spread spectrum operation on the input baseband signal;

b, and then continue to perform channel-divided Walsh despreading on the despread signal, and separate the signals of each channel of the user;

c multiplying the separated bit streams of each channel with a conjugate signal of an output estimate value of a pilot estimator A to counteract the influence of multipath fading;

d, the multiplied bit stream is sent to the judger for judgment;

e, using the other two pilot estimators 8, C and the number of users, channel types, and service types provided by the system to construct a threshold required for judgment, and then judge the input bit stream to be judged;

f. Multiply the determined bit stream by the estimated value output by the pilot estimator A to reconstruct the influence of multipath fading and the subsequent reconstruction of the signal.

In said step d, when making a judgment, the following steps are further included:

dl , obtain parameters such as the number of users and channel type or service type from the system;

02. Determine the channel type or service type information;

d3, adjust the coefficient K2 in the formula ^^H according to the information of step d2;

d4, determine the current number of users;

d5, adjust the coefficient K1 in the formula in step d3 according to the number of users.

In the step d3, when adjusting the coefficient K2, the higher the channel rate, the higher the value of K2, and the K2 corresponding to the convolutional coding channel is greater than that of the Turbo coding type channel.

When adjusting the coefficient, K2 is determined first, and then K1 is determined. The value of K2 depends on the increase of the channel rate. The change of height and coding type increases roughly linearly. During specific operation, K1 can be temporarily set to about 1 to 3, and then simulate and test the optimal value of K2 under different channel rates and different coding methods.

A double-threshold decision can be used when making a decision in step e, that is, when the energy of the bit to be decided is less than the threshold 1 (T1), the decision is 0; otherwise, when the energy of the bit to be decided is greater than the threshold 2 (T2), Also the judgment is 0; otherwise the judgment is 1 or -1. In the step d3, it can also be adjusted according to the formula ^^W mm ^- ^, wherein K3 is a confidence factor, which is a number not less than 1.

When making adjustments,

First obtain the output signal of the pilot estimator C, and then according to the formula Energy[PilotChip _k ]

E C =

_k

and the formula

Bas = bE_C

Calculating the benchmark Bas of E-B, judging whether E-B is greater than Bas, if greater, indicating that the bit energy is stronger, then reducing K3;

Then judge whether K3 is reduced to a value less than 1, if less than 1, then set K3 = 1 _; if Ε_B is less than Bas, indicating that the bit energy is weak, then maintain or increase K3;

Then judge whether K3 exceeds the maximum value MAX_K3, where MAX_K3 is an integer greater than 1.

The multi-user interference cancellation unit of the present invention includes a despreading unit, three pilot estimators A, B, C, a first multiplier, a conjugate device, a decision device, and a second multiplier, and the despreading unit decomposes the input signal spread, separate the useful signal of the user from the total signal, send the despread signal to three pilot estimators on the one hand, and send the despread bit stream to the first multiplier on the other hand, guide The output signal of the frequency estimator A is conjugated by the conjugator, and the first multiplier multiplies the despread bit stream and the signal conjugated by the conjugator; the decision unit uses the output of the pilot estimator 8 and C The signal and system information judge the bit stream to be judged output by the first multiplier; the second multiplier multiplies the output signal of the judger and the signal output by the pilot estimator A again, and the result of the multiplication is the solution of the subsequent stage input to the tuning module. Figure overview

Figure 1 is a schematic block diagram of the existing receiver principle.

Fig. 2 is a schematic diagram of an existing parallel interference cancellation structure for multi-user detection.

Fig. 3 is a schematic diagram of the structure of the interference cancellation unit of the present invention.

Fig. 4 is a schematic flow chart of the interference elimination method of the present invention.

Fig. 5 is a schematic diagram of the process of the interference elimination method (changing the decision threshold according to different situations) of the present invention. Fig. 6 is a schematic diagram of the demodulation performance curve of the receiver under different decision thresholds after using the method of the present invention.

Fig. 7 is a schematic diagram of the judgment (how to reduce the error rate of judgment) process in the interference elimination method of the present invention. Fig. 8 is a schematic diagram of the judgment (how to measure the confidence of the judgment) process in the interference elimination method of the present invention. BEST MODE FOR CARRYING OUT THE INVENTION

Figure 1 illustrates the basic principle of the receiver, wherein the radio frequency module converts the received radio frequency signal into a baseband signal and sends it to the demodulation module for demodulation, and the demodulated symbol stream is then sent to the decoder for decoding. Usually, the RF processor includes a mixer and local oscillator, which can convert the RF signal into an intermediate frequency signal, and then further convert it into a baseband signal.

The demodulation module includes a RAKE receiving module to search for the strongest k multipath signals; a despreading module uses the mixed PN spreading sequence corresponding to each user to despread each path signal of each user; finally, the The despread symbol stream is multiplied by the Walsh function sequences corresponding to different channels, and the obtained bit stream can be sent to the decoder for decoding.

The decoder selects the corresponding decoding method according to different coding methods of the signal, for example: the fundamental channel (Fundamental channel) uses the convolutional coding method, and the corresponding decoding method is Viterbi

(Viterbi) decoding; and when the supplemental channel (Supplemental channel) uses Turbo coding, the corresponding decoding method is Turbo decoding.

Fig. 2 schematically shows an existing parallel interference cancellation (PIC: Parallel Interference Canceller) structure for multi-user detection. The structure includes M-order multi-user detection processing (M>1), and each stage includes N ICUs, where N is the number of users. The N ICUs are arranged in parallel to perform interference cancellation at the same time, instead of first selecting the strongest user for interference cancellation like serial interference cancellation. Every The delay length of the first-order delay units 201-1 to 201-M is equal to the processing delay of the corresponding stage, so as to ensure the synchronization of the two input signals of the subtractor 204-n (n is 1~-M).

It is worth noting that the interference cancellation structures of the first and last orders are slightly different from those of the other orders. The input signal of the first stage is obtained from the output of the matched filter 206, and the input signals of other stages are the outputs of the subtractor of the previous stage, that is, the signal after interference cancellation. The last level of ICU 2-M-1 to 2-M-N only needs to perform interference signal reconstruction. After all stages of multi-user detection are over, the final output signal An is the reconstructed signal by subtracting the interference of all other users except user n from the total signal (n is 1~-N). Theoretically, if the signals of all users can be accurately reconstructed, even in the case of multiple users, its demodulation performance is the same as that of one user.

Please refer to FIG. 3 first, which is a schematic structural diagram of the interference cancellation unit (ICU) in FIG. 2, that is, the structure of the interference cancellation unit adopted in the present invention. The unit includes a despreading unit, three pilot estimators A, B, C, a first multiplier, a conjugate device, a decision device, and a second multiplier. The despreading unit despreads the input signal, and the user's useful The signal is separated from the total signal, and the despread signal is sent to the three pilot estimators on the one hand, and the despread bit stream is sent to the first multiplier on the other hand, the output of the pilot estimator A The signal is conjugated by the conjugator, and the first multiplier multiplies the despread bit stream and the signal conjugated by the conjugate device; the decision unit uses the output signal of the pilot estimator 8, C and the system information A multiplier outputs the bit stream to be judged; the second multiplier multiplies the output signal of the judger and the signal output by the pilot estimator A again, and the multiplied result is the input of the subsequent demodulation module.

However, there is only one pilot estimator in the structure of the traditional ICU, and the signal of the pilot estimator is not only used for weighting the despread signal and reconstructing the received signal, but also used as a reference signal for the decision unit. However, since the essence of weighting the despread signal and reconstructing the received signal is to cancel and reconstruct the multipath fading effect in the received signal, it is required that the cumulative length of the pilot estimator should not be too long, and the delay should not be too large, so that the obtained The multipath fading component can be included in the pilot frequency estimation with sufficient accuracy; however, such pilot frequency estimation is not suitable for the reference signal of the decision maker. Therefore, dedicated pilot estimators B and C are added to the ICU structure provided by the present invention, which are used as reference signals for the decision unit, and pilot estimator A is dedicated to weighting the despread signal and reconstructing the received signal.

FIG. 4 is a flow chart of the interference elimination method of the present invention. The method of the present invention is as follows: the first step 401 uses the Despread the input signal according to the mixed PN spreading sequence of the user, separate the useful signal of the user from the total signal, and send the despread signal to the pilot estimator AC. In the second step 402, the channel-divided Walsh despreading is performed on the signal after PN despreading, and the signals of each channel of the user are separated. Both steps 401 and 402 are performed in the despreading unit. Steps 404, 408, and 410 extract pilot frequency estimates A, B, and C respectively according to different pilot frequency estimation methods. In step 412, the separated bit streams of each channel are multiplied by the conjugate signal of the output signal of the pilot estimator A to offset the influence of multipath fading, and the multiplied bit streams are sent to the decision unit for decision. Step 416 utilizes the pilot frequency estimation 8, C and the system information provided in step 414 to construct the threshold required for judgment, and then judges the input bit stream to be judged. Step 418 multiplies the judged bit stream by the pilot estimate A to reconstruct the influence of multipath fading, and then the subsequent reconstruction process of the signal, including walsh spread spectrum, mixed PN sequence spread spectrum, etc., will not be described in detail . The outputs of the pilot estimator B and the pilot estimator C in FIG. 3 and the system information are the inputs required by the decider to make a decision, and their detailed functions will be described below.

The role of the decider is to judge the input bit stream into a sequence of 1 and -1. The simplest judgment method is: if the amplitude of a certain bit is greater than 0 level, it is judged as 1; otherwise, it is judged as -1. Considering the influence of thermal noise and multipath fading, this decision method is obviously very imprecise. In order to judge more accurately and avoid introducing additional interference as much as possible, it is necessary to introduce a threshold for the judgment process. The new judgment method is: if the magnitude of a certain bit is greater than the threshold, it will be judged as 1; if the magnitude is smaller than the negative threshold, the judgment will be -1; in other cases, it will be judged as 0 level. This method takes into account that when the bit amplitude is very weak, the reliability of the decision of the bit is relatively low, instead of judging it as a non-zero value, it is better to judge it as 0 to avoid introducing possible additional interference. Also: Due to the large fluctuation of bit stream energy caused by multipath fading, the use of a fixed decision threshold will obviously not bring accurate decision results, and because the size of the bit energy is unpredictable, it is difficult to specify a fixed threshold. Achieved. In consideration of the above factors, the threshold used in the present invention can be obtained by the following formula:

Threshold = K\ - K2 - E_B ₍₁₎ where EJ3 is the energy value of the output signal of the pilot estimator B, K1 depends on the number of users, K2 depends on the channel type or service type of the corresponding user, both K1 and K2 is a non-negative number. For example: If the user is using a 153.6kbps supplementary channel, its K2 is greater than the 9.6kbps basic channel. Both K1 and K2 need to be obtained through simulation and real-world testing. Fig. 6 is the demodulation performance curve obtained by simulation under different decision thresholds for different numbers of users, and the channel used is a basic channel of 9.6 kbps. It can be found that: when the number of users is different, the optimal decision threshold is also different. That is to say, K1 changes as the number of users changes. The optimal K1 and K2 must be summarized on the basis of a large amount of simulation data and real-world test data.

Fig. 5 is a schematic diagram of the process of the interference elimination method (changing the decision threshold according to different situations) of the present invention. The process is as follows: the first step 502 is to obtain parameters such as the number of users and channel type or service type from the system, and these parameters are easy to obtain; the second step 503 is to determine what the channel type or service type is, for example, it is a basic channel or a supplementary channel channel, what is the channel rate and what is the encoding type, etc. The third step 504 adjusts the coefficient K2 according to the information in the second step 503, for example: the higher the channel rate, the higher the value of K2; the K2 corresponding to the convolutional coding channel is greater than that of the Turbo coding type channel. Generally speaking, the value of K2 shows a roughly linear upward trend with the increase of the channel rate and the change of the encoding type (from convolutional encoding to Turbo encoding). For specific operations, K1 can be temporarily set to about 2 to 3, and then The optimal value of K2 is simulated and tested under different channel rates and different encoding methods. The optimal value has little relationship with the number of users, so the number of users in the test can vary from 10 to 40. After K2 is determined, the fourth step 505 will determine the current number of users, and the fifth step 506 will adjust K1 according to the number of users. For example: Referring to Figure 6, when the number of users is between 10 and 20, the best K1 is 1; when the number of users is 30, the best K1 is 2; when the number of users is 40, the best K1 is 4 or 5, the value range of K1 can be easily determined through simulation, too large or too small is not good. Figure 6 is only the simulation result in a specific environment, to obtain the best Kl suitable for various environments, a lot of simulation and real-world testing are needed. In different communication environments, the optimal value of K1 will be shifted, but the shift will not be very large, so an equilibrium point can be selected among these optimal values, so that it can basically take into account various communication environments, Such an equilibrium point is the optimal K1 value for the current number of users. During specific implementation, a table of K1 and K2 can be made according to the test data, and the system can check the corresponding K1 and K2 according to the real-time system information (parameters such as the number of users, channel type, coding type, etc.).

E-B in the above formula (1) uses the energy value of the pilot estimator B. There is only one pilot estimator A in the traditional ICU. Since the pilot estimator needs to weight the bit stream before and after the decision to offset and reconstruct the influence of multipath fading, the delay of the pilot estimator cannot be too long , the cumulative length cannot be too large, but the variance of the signal output by such a pilot estimator must be large, and the energy The fluctuation is also severe, which is not suitable for generating a decision threshold. Therefore, a pilot estimator B needs to be added to complete this work. The difference between pilot estimator B and pilot estimator A is that the cumulative length of pilot estimator B is longer, so the variance of its output signal is small, the energy fluctuation is small, and its energy envelope basically conforms to the bit to be decided The energy envelope of the flow is more suitable for generating the decision threshold. It can be known from the simulation experiment that the demodulation effect after using the pilot estimator B to generate the decision threshold is obviously improved compared with using the pilot estimator A only and using the pilot estimator A and B in combination.

When judging, if the bit energy is weak, it means that the reliability of the judgment is low; in the same way, if the bit energy is very strong and exceeds a certain limit, the energy should be regarded as an abnormality caused by interference Obviously, the judgment reliability of this bit is also low. Based on this point of view, in order to further reduce the additional interference caused by judgment errors, the present invention adopts a double-threshold judgment. That is: when the energy of the bit to be judged is less than the threshold 1, or the energy of the bit is greater than the threshold 2, the bit should be judged as 0; in other cases, the judgment is 1 or -1. Threshold 2 can be obtained by adding a factor on the basis of threshold 1, namely P:

71 = ^1, - ΛΓ2, E_B ( ₂ ) T2 = a - T\ (3)

Wherein, a is a number greater than 1, and the optimal value of a can be obtained through simulation and real-world testing (when a is infinite, it is equivalent to only the decision threshold T1). Fig. 7 is a judgment process based on double thresholds. The decision principle is as above, namely: when the energy of the bit to be decided is less than T1, the decision is 0 _; otherwise, when the energy of the bit to be decided is greater than T2, the decision is also 0; in other cases, the decision is 1 or -1. The double-threshold decision can more effectively reduce errors in the decision process, avoid error accumulation in the multi-stage processing process, and reduce the introduction of additional interference.

In the code division multiple access communication system, the reverse pilot channel is transmitted together with the traffic channel modulation, and the air channel fading experienced is the same, therefore, the shape of the energy envelope of the pilot channel and the traffic channel is also basically the same , when the energy of the traffic channel becomes stronger, the energy of the pilot channel also becomes stronger. Then, it can be imagined that since the decision threshold of the above method is generated by using the signal of the pilot estimator, the envelope shape of the decision threshold is basically the same as the energy envelope shape of the traffic channel. In other words, the decision threshold The envelope shape is basically the same as the energy envelope shape of the bitstream to be decided. In this way, when the energy of the bit stream is strong, because the decision threshold also increases, the bit stream is judged The probability of being 0 will also be greater. According to the above statement, when the energy of the bit to be judged is strong, the reliability of the judgment is high; otherwise, the reliability of the judgment is low. Considering that the energy change of the bit stream is continuous and gradual, there is the following idea, when the energy of the bit stream is strong, reduce the probability of the bit being judged as 0, and in the time period of the bit stream energy is weak, Then maintain or even increase the probability that a bit is judged to be 0. This can further improve the accuracy of the decision.

In order to realize the above viewpoint, a confidence factor can be added to the decision threshold, and structurally, a pilot estimator C can be added in the interference elimination unit. The output of the pilot estimator C is the long-term average value of the energy of the pilot channel, which is used to measure the strength of the output energy of the pilot estimator B. The implementation structure of pilot estimator C is different from that of pilot estimators A and B, which can be expressed by the following formula-

Energy[PilotChip _k ]

E C = ^

- k (4 ) is the long-term average of the energy of the pilot channel from the beginning of channel establishment to the present. The reason why the long-term average value is used is that it is relatively stable, and is less affected by multipath fading, and the energy fluctuation is also small, so it is more suitable as a benchmark for measuring the output signal of the pilot estimator B. The benchmark for measurement is obtained by the following formula:

Bas = b.E—C (5) where b is a factor less than 1 and greater than 0, and its optimal value needs to be obtained through simulation and real-world testing; E—C is the output value of the pilot estimator C.

Therefore, the formula (1) of the decision threshold can be updated as:

Threshold = Kl - K2 - K3 - E_B ( ₆ ) where K3 is the confidence factor, which is a number not less than 1. The adjustment process is shown in Figure 8: first obtain the output signal of the pilot estimator C, and then according to the formula ( 4) Calculate the benchmark Bas of E_B, and judge whether E-B is greater than Bas. If it is greater, it indicates that the bit energy is stronger, and K3 is reduced, which is equivalent to reducing the judgment threshold and reducing the probability that the bit stream to be judged is judged to be 0. Then judge whether K3 is reduced to a value less than 1, if it is less than 1, then set K3 = 1; if E-B is less than Bas, indicating that the bit energy is weak, then maintain or increase K3, which is equivalent to increasing the decision threshold, increasing Determine the probability that the bit stream to be judged is judged to be 0, and then judge whether K3 exceeds the maximum value MAX_K3, and if so, make

K3=MAX—K3. Wherein, MAX_K3 is an integer greater than 1, and the specific value can be obtained through simulation and Field testing OK.

In the entire interference cancellation technology, whether the useful signal of each user can be accurately reconstructed is the key to the effectiveness of the interference cancellation technology, and among them, whether the despread bit stream in the ICU can be accurately judged is whether the useful signal can be accurately reconstructed Base. The present invention proposes a series of methods around how to reduce the error rate of judgment. Compared with other interference cancellation technologies, these methods are effective and easy to implement, and the required input information is also easy to obtain, which improves the demodulation performance of the receiver It is more obvious, which is beneficial to increase the system capacity and the system coverage radius. ·

The interference elimination method and unit of the present invention are mainly applied to a receiver with a multi-user detector, which can improve the demodulation performance of the base station and increase the capacity.

The elimination method and the elimination unit proposed by the present invention are not only applicable to the parallel interference elimination structure, but also applicable to the serial interference elimination structure, and there is no need for much change. The present invention can also be applied to other digital communication systems or analog communication systems through proper modification, as long as the mobile station of the system also transmits the pilot channel and conforms to the IS-665 standard. Industrial applicability

Since the present invention adopts the above-mentioned technical solution, compared with the prior art, it overcomes that when it is less than the threshold, the reliability of the received signal is poor, it is very likely to misjudgment and misjudgment, and, in multi-stage processing, the accumulation of errors The disadvantage of causing a rapid decline in performance. The present invention performs more accurate control and processing on signal judgment and signal recovery, increases the accuracy of interference elimination, thereby improving system performance; further increases system capacity and alleviates the impact of "near-far effect" on system performance; adopts After the receiver after the multi-user detector constituted by the interference elimination unit of the present invention, the demodulation performance of the base station can be improved and the capacity can be increased.

Claims (6)

Rights request 1. A method for multi-user interference cancellation, characterized in that the method comprises the following steps: a, using a despreading unit to perform a decomplex spread spectrum operation on an input baseband signal; b, and then continue to perform channel-divided Walsh despreading on the despread signal, and separate the signals of each channel of the user; c multiplying the separated bit streams of each channel with a conjugate signal of an output estimate value of a pilot estimator A to counteract the influence of multipath fading; d, the multiplied bit stream is sent to the judger for judgment; e, using the other two pilot estimators 8, C and the number of users, channel types, and service types provided by the system to construct a threshold required for judgment, and then judge the input bit stream to be judged; f. Multiply the determined bit stream by the estimated value output by the pilot estimator A to reconstruct the influence of multipath fading and the subsequent reconstruction of the signal. 2. The method for multi-user interference cancellation according to claim 1, characterized in that, in step d, when making a decision, further comprising the following steps: dl, obtain parameters such as the number of users and channel type or service type from the system; d2 , determine the channel type or service type information; d3, adjust the coefficient K2 in the formula ^^^^ ^ ' ^ ^^ according to the information of step d2; d4, determine the current number of users; d5, adjust the coefficient K1 in the formula in step d3 according to the number of users. 3. The method for multi-user interference cancellation according to claim 2, characterized in that: in the step d3, when adjusting the coefficient K2, the higher the channel rate, the higher the value of K2, and the convolutional coding channel corresponds to The K2 is larger than the channel of Turbo coding type. 4. The method for multi-user interference cancellation according to claim 3, characterized in that: when adjusting the coefficients, first determine K2, and then determine K1, and the value of K2 varies with the increase of the channel rate and the change of the coding type It rises roughly linearly. In specific operations, you can temporarily set K1 to about 2 to 3, and then simulate Truth and test the optimal value of K2 under different channel rates and different encoding methods. 5. The method for multi-user interference cancellation according to claim 2, characterized in that: a double-threshold decision can be used when making a decision in said step e, that is, when the energy of the bit to be decided is less than the threshold 1 (T1) , the judgment is 0; otherwise, when the energy of the bit to be judged is greater than the threshold 2 (T2), the judgment is also 0; in other cases, the judgment is 1 or -1. 6. The method for multi-user interference cancellation according to claim 2, characterized in that: in the step d3, it can also be adjusted according to the formula ^ ^ -mm^-, wherein K3 is a confidence factor, which is not less than The number of 1. 7. The method for multi-user interference cancellation according to claim 6, characterized in that: when making adjustments, First obtain the output signal of the pilot estimator C, and then according to the formula Energy[PilotChip _k ] E C = ^ one k and the formula Bas = b - E_C Calculating the benchmark Bas of E_B, judging whether E_B is greater than Bas, if greater, indicating that the bit energy is stronger, then reducing K3; Then judge whether K3 is reduced to a value less than 1, if less than 1, then set K3 = 1; if Ε—Β is less than Bas, indicating that the bit energy is weak, then maintain or increase K3; Then judge whether K3 exceeds the maximum value MAX_K3, where MAX_K3 is an integer greater than 1. 8. A multi-user interference cancellation unit, characterized in that: the interference cancellation unit includes a despreading unit, three pilot estimators A, B, C, a first multiplier, a conjugate unit, a decision unit, and a second multiplier The despreading unit despreads the input signal, separates the useful signal of the user from the total signal, sends the despreaded signal to the three pilot estimators on the one hand, and sends the despreaded signal to the three pilot estimators on the other hand. The bit stream is sent to the first multiplier, the output signal of the pilot estimator A is conjugated by the conjugator, and the first multiplier multiplies the despread bit stream and the signal conjugated by the conjugator; Using the output signal of the pilot estimator 8, C and the system information to judge the bit stream to be judged output by the first multiplier; the second The multiplier multiplies the output signal of the decision device and the signal output by the pilot estimator A again, and the multiplied result is the input of the subsequent demodulation module.